Ethereal Digital Hiring! Full Time Artificial Intelligence Engineer - Senior in Federal Territory - Ricebowl

Employement Type : 6 Month Contract

Project: Bank Negara Malaysia

Note: Open for local Malaysian candidate only

Job Requirement

• Own ambiguous problems, not just scoped projects. Take a problem the bank brings us before it has a defined solution and decide its shape: what to build, what to buy, what to leave out, and what “good enough to ship and operate” means for it. Turn it into a scoped project and a defensible system design the rest of the team can build against.
• Be the technical anchor on the project. On the projects you run, you are the most senior engineer in the room and the technical point of reference the rest of the team builds around. Define the milestones, own the decision log, and keep the design coherent as engineers rotate on and off, because in our swarming model the team changes before the project ships and the continuity lives in your documentation and your head. This is a senior engineering role, not a lead or management one. You anchor the work by being the person others rely on technically, not by holding a title.
• Design moderately complex systems independently. You own your system designs without needing a more senior engineer to sign off before you build. Reason about latency budgets, fault tolerance, graceful degradation, state management, and how the system behaves when an AI component is slow, wrong, or replaced by a provider upgrade. Decide the system’s shape before anyone writes code, and justify it.
• Run the full lifecycle, including governance. Take the system from design through evaluation, MRM coordination, and production rollout. You own the system the model runs in, including its documentation burden for Model Risk Management (MRM) review: prompts, retrieval sources, tool surfaces, guardrails, evaluation results, and known failure modes. You are not the owner of the model’s behaviour or its governance package. That’s the ML Engineer’s role. You own the system the model runs in, and you coordinate the governance process for that system rather than waiting to be told what it needs.
• Design observability in from day one. Decide what to log, measure, and trace before the system ships. Balance offline evals against online metrics and human review. Catch the subtle regressions that junior and mid-level engineers tend to miss: drift across model upgrades, retrieval-index refreshes that quietly degrade quality, latency creep. Know which signals mean “block the release.”
• Pick the right pattern, and track what’s actually possible. Choose among the orchestration patterns the lab uses (retrieval, structured output, tool use, agentic loops) based on the problem, not the trend. Reason about prompt injection, hallucination, and data leakage as engineering problems with engineering defences. Track what frontier models can and can’t do well enough to tell a stakeholder when AI is the wrong tool.
• Set the design and review bar. Your code reviews and design feedback raise the team’s standard by example. Mentor junior engineers, conduct levelling interviews, and push the team’s practices forward: testing, observability, decision logs, and the kinds of design choices that compound over years.
• Push back with evidence. Bring real usage data and eval results to the table. Disagree productively with stakeholders when a requirement doesn’t match what AI can do or what the system can sustain in production. Optimise for the bank’s outcome, not just closing the ticket.
• Collaborate across functions as the project’s technical voice. Translate between business stakeholders, the AI/ML Engineering team, validators, and downstream production teams. Represent what the system does, what it can’t, and where the design has limits, in plain language to a stakeholder and in technical detail to a reviewer.

What we're looking for

• Experience. Minimum 4 years of professional software engineering experience, including building and shipping production systems that integrate AI components (LLMs, classical ML, or both), and at least one where you owned the system design rather than just the implementation. Time spent as the senior engineer anchoring a project with more than one engineer, holding the design together as others come and go. Bachelor’s in Computer Science, Software Engineering, or a related technical field; equivalent experience considered.
• Software engineering depth. Deep fluency in our stack. TypeScript with React on the frontend, and Python with FastAPI on the backend. You’ve built systems real users or internal consumers depend on, and you’ve maintained them long enough to live with your own design decisions. You set the standard for style, maintainability, and testability rather than just meeting it. Comfortable with Git, code review, CI, and reproducible workflows. Coding-assistant fluency is assumed, not differentiated.
• System design. You design moderately complex systems independently. Given an ambiguous problem, you decide the system’s shape: components, contracts, state management, failure modes, and observability, and you can defend both the choices and the things you chose not to do. You reason about how the system degrades when an AI component is slow, wrong, or upgraded by its provider, and you design for that before writing code.
• AI literacy as a bar requirement. You’ve built production-grade systems with LLMs: RAG, structured output, tool use, memory, agentic patterns, and you know the range well enough to pick the right one rather than reaching for the same one each time. You consume and integrate ML models built by the AI/ML Engineering team. You treat prompt injection and data leakage as engineering concerns with engineering defences. You track frontier model capabilities closely enough to know when a use case doesn’t fit one. Eval methodology may not be your speciality, but you know what a good eval looks like and you build evaluation into the system.
• Production ownership. You’ve owned live systems, not just shipped them. You’ve designed observability in from the start, run incidents, written postmortems, and shipped the follow-ups. You catch the subtle regressions: drift, index degradation, latency creep. You don’t trust offline metrics to predict live behaviour, and you know which signals justify holding a release.
• Technical leadership without management. You’ve been the technical anchor on a project: the person who held the design context, owned the decision log, and kept teammates aligned as they cycled on and off. You mentor more junior engineers and raise the bar through review rather than decree. This is a senior individual-contributor role. It is not a people-management role, and the leadership we’re describing is technical.
• Communication. You write design docs and decision logs that someone joining the project mid-flight can actually follow. You can explain what your system does and where it fails, in plain language to a stakeholder and in technical detail to a reviewer. You disagree productively, with evidence, when requirements don’t match what the system can sustain.